skey.git - Linux Kernel Module for the Keyboard of the Omnicom

Age	Commit message (Collapse)	Author
2017-12-28	finished the keymap	knolax

2017-12-28	added systemd unit to initialize pin modes and load the kernel module on ↵	knolax
	startup, yet to be tested
2017-12-28	removed entry for device, the gpio settings for all pins except bcmpin 15, ↵	knolax
	which is set to alt0 or RXD mode. messing around with dtoverlays is really a waste of time so i'm moving on to a systemd daemon that just runs the gpio utility to change the pinmodes, and then load the module
2017-12-28	fixed the target for install	knolax

2017-12-28	fixed the format for compatible tag, still doesn't explain why the gpios ↵	knolax
	aren't being applied right
2017-12-28	added install target to makefile, not sure if i have to turn the *.ko file ↵	knolax
	into a gunzip, so testing this
2017-12-28	after some testing, the issues with bcmpins 4 and 15 were indeed that they ↵	knolax
	were pull-up, this means that the issue lies with the device tree overlay not working as intended. two properties have been added to hopefully make this overlay load as intended. the compatible field has been uncommented as the driver is stable enough to be loaded automatically.
2017-12-27	undid 88f06d, this was already done in the production code but i forgot to ↵	knolax
	update it for git.
2017-12-27	made the DTO more proper, the compatible property is still commented out	knolax

2017-12-27	set pin 4 and 15 to have low values to, this might not actually run.	knolax

2017-12-27	after switching the position of 4 in column_pins[10] the problem persisted, ↵	knolax
	i have added lines at the end of init_skey() that explictly set those two pins as input, this is to help debug the issue]
2017-11-14	reverted the inverted row pins, the issue was not one of ↵	knolax
	active_low/active_high, but an incomplete transition to using column pins for input and row pins for output. everything works except for columns 7 and 9(pins 15 and 4) which are always held detected as hgih. testing with the /sysfs interface has shown that pin 4 is electrically sound so it must be an issue with the driver# Please enter the commit message for your changes. Lines starting
2017-12-23	fixed typo	knolax

2017-12-23	fixed typo	knolax

2017-12-23	fixed testing setup	knolax

2017-12-23	hopefully fixed setting pins as opposite of what they are, thereby fixing ↵	knolax
	the issue of the iutput pins not working
2017-12-23	making a test, gpio output seems unresponsive at all	knolax

2017-12-23	switched the output pin states as they might be considered active_low in the ↵	knolax
	legacy gpio interfacd
2017-12-23	fixed typo	knolax

2017-12-23	changed the function calls to the linux gpio interface to their cansleep ↵	knolax
	equivalents
2017-12-20	added a small sleep between rpin state changes	knolax

2017-12-20	fixed typo	knolax

2017-12-20	reversed the roles of the row and column pins to account for an error in the ↵	knolax
	prototype board.
2017-12-19	changed tick rate of update thread for testing purposesHEAD master	knolax

2017-12-19	added target in maekfile for making the dtbo	knolax

2017-12-19	actually added the pinsthread	knolax

2017-12-19	added device tree overlay to make pins pull down`	knolax

2017-12-13	reenabled gpio testing	knolax

2017-12-13	confirmed that adding the metadata allows X11 to autodetect skey_dev	knolax

2017-12-13	removed __initdata from all the variables because none of them are actually ↵	knolax
	only used during init
2017-12-13	added more metadata to skey_dev, also added __initdata to some variables	knolax

2017-12-08	fixed a typo, has it emmit capslock every second	knolax

2017-12-08	added code to register a device with input.h	knolax

2017-12-08	aded keymap header and custom macros for custom keys	knolax

2017-12-08	added a counter that printks every second, the previous kernel panics may ↵	knolax
	have been caused by excessive use of printk.
2017-12-08	changed test mode warning to only emit once during the thread	knolax

2017-12-08	added test mode that doesn't use GPIO so i can test on an Virtual Machine on ↵	knolax
	my main computer
2017-12-07	changed the update function from being timer based to thread based so that ↵	knolax
	it can sleep. it may be because of the fact that it couldn't sleep that it was causing kenel panics
2017-12-07	added functions to explicitly set pins as input or output do to kernel ↵	knolax
	panics when trying to use gpio_set_value()
2017-12-07	added basic keypress scanning functionality to the update function	knolax

2017-12-07	initial commit	knolax